3D Part Segmentation on ShapeNetPart (val)
87.82mIoUPIC++
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| PIC++variant=-Sep, training_protocol=full fine-tuning2024.04 | 87.82 | — | — | |
| PIC++variant=-Cat, training_protocol=full fine-tuning2024.04 | 87.31 | — | — | |
| PointMLPTraining Strategy=Multi-Dataset Joint Training2024.04 | 86.32 | — | — | |
| ACTTraining Strategy=Multi-Dataset Joint Training2024.04 | 85.98 | — | — | |
| PIC++variant=-Cat, training_protocol=scratch2024.04 | 85.32 | — | — | |
| PCMambaTraining Strategy=Multi-Dataset Joint Training2024.04 | 83.89 | — | — | |
| PCTTraining Strategy=Single Dataset Supervised Training2024.04 | 83.71 | — | — | |
| MAMBA3DTraining Strategy=Multi-Dataset Joint Training2024.04 | 83.51 | — | — | |
| PIC++variant=-Sep, training_protocol=scratch2024.04 | 83.26 | — | — | |
| PTv1Training Strategy=Multi-Dataset Joint Training2024.04 | 83.24 | — | — | |
| PTv2Training Strategy=Multi-Dataset Joint Training2024.04 | 83.18 | — | — | |
| CurveCloudNetType=Curve2023.03 | 83.1 | — | — | |
| PointNeXtTraining Strategy=Multi-Dataset Joint Training2024.04 | 82.92 | — | — | |
| CurveNetType=Point2023.03 | 82.8 | — | — | |
| PointNextType=Point2023.03 | 82.8 | — | — | |
| PTv3Training Strategy=Multi-Dataset Joint Training2024.04 | 82.45 | — | — | |
| PointDiffTraining Strategy=Multi-Dataset Joint Training2024.04 | 82.14 | — | — | |
| PCTTraining Strategy=Multi-Dataset Joint Training2024.04 | 82 | — | — | |
| PICvariant=-Sep2024.04 | 81.72 | — | — | |
| MinkowskiNetType=Voxel2023.03 | 81.1 | — | — | |
| PointMLPType=Point2023.03 | 80.9 | — | — | |
| PointNetTraining Strategy=Single Dataset Supervised Training2024.04 | 80.12 | — | — | |
| PointNet++Type=Point2023.03 | 80.1 | — | — | |
| Cylinder3DType=Voxel2023.03 | 79.6 | — | — | |
| SphereFormerType=Voxel2023.03 | 79.5 | — | — | |
| PointM2AETraining Strategy=Multi-Dataset Joint Training2024.04 | 79.46 | — | — | |
| PointNetTraining Strategy=Multi-Dataset Joint Training2024.04 | 79.07 | — | — | |
| DGCNNTraining Strategy=Multi-Dataset Joint Training2024.04 | 78.58 | — | — | |
| PointLoRATraining Strategy=Multi-Dataset Joint Training2024.04 | 78.47 | — | — | |
| DGCNNTraining Strategy=Single Dataset Supervised Training2024.04 | 78.23 | — | — | |
| I2P-MAETraining Strategy=Multi-Dataset Joint Training2024.04 | 77.68 | — | — | |
| PICvariant=-Cat2024.04 | 70.5 | — | — | |
| Copy2024.04 | 23.18 | — | — | |
| ACTYear=2023, Learning Protocol=with Self-Supervised Representation Learning (FULL)2023.09 | — | 84.66 | 86.14 | |
| DGCNNYear=2019, Learning Protocol=Supervised Learning Only2023.09 | — | 82.33 | 85.2 | |
| Point-BERTYear=2022, Learning Protocol=with Self-Supervised Representation Learning (FULL)2023.09 | — | 84.11 | 85.6 | |
| Point-MAEYear=2022, Learning Protocol=with Self-Supervised Representation Learning (FULL)2023.09 | — | — | 86.1 | |
| Point-RAEYear=2023, Learning Protocol=with Self-Supervised Representation Learning (FULL)2023.09 | — | 84.71 | 86.28 | |
| PointMLPYear=2022, Learning Protocol=Supervised Learning Only2023.09 | — | 84.6 | 86.1 | |
| PointNetYear=2016, Learning Protocol=Supervised Learning Only2023.09 | — | 80.39 | 83.7 | |
| PointNet++Year=2017, Learning Protocol=Supervised Learning Only2023.09 | — | 81.85 | 85.1 | |
| TransformerYear=2017, Learning Protocol=with Self-Supervised Representation Learning (FULL)2023.09 | — | 83.42 | 85.1 |